We survey 146 papers analyzing "bias" in NLP systems, finding that their motivations are often vague, inconsistent, and lacking in normative reasoning, despite the fact that analyzing "bias" is an inherently normative process. We further find that these papers' proposed quantitative techniques for measuring or mitigating "bias" are poorly matched to their motivations and do not engage with the relevant literature outside of NLP. Based on these findings, we describe the beginnings of a path forward by proposing three recommendations that should guide work analyzing "bias" in NLP systems. These recommendations rest on a greater recognition of the relationships between language and social hierarchies, encouraging researchers and practitioners to articulate their conceptualizations of "bias"---i.e., what kinds of system behaviors are harmful, in what ways, to whom, and why, as well as the normative reasoning underlying these statements---and to center work around the lived experiences of members of communities affected by NLP systems, while interrogating and reimagining the power relations between technologists and such communities.

Language (Technology) is Power: A Critical Survey of "Bias" in NLP

We survey 146 papers analyzing "bias" in NLP systems, finding that their motivations are often vague, inconsistent, and lacking in normative reasoning, despite the fact that analyzing "bias" is an inherently normative process We further find that these papers' proposed quantitative techniques for measuring or mitigating "bias" are poorly matched to their motivations and do not engage with the relevant literature outside of NLP Based on these findings, we describe the beginnings of a path forward by proposing three recommendations that should guide work analyzing "bias" in NLP systems These recommendations rest on a greater recognition of the relationships between language and social hierarchies, encouraging researchers and practitioners to articulate their conceptualizations of "bias"---ie, what kinds of system behaviors are harmful, in what ways, to whom, and why, as well as the normative reasoning underlying these statements---and to center work around the lived experiences of members of communities affected by NLP systems, while interrogating and reimagining the power relations between technologists and such communities

/pdf/language-technology-is-power-a-critical-survey-of-bias-in-40xzysfmb0.pdf

Hate speech detection is a challenging problem with most of the datasets available in only one language: English. In this paper, we conduct a large scale analysis of multilingual hate speech in 9 languages from 16 different sources. We observe that in low resource setting, simple models such as LASER embedding with logistic regression performs the best, while in high resource setting BERT based models perform better. In case of zero-shot classification, languages such as Italian and Portuguese achieve good results. Our proposed framework could be used as an efficient solution for low-resource languages. These models could also act as good baselines for future multilingual hate speech detection tasks. We have made our code and experimental settings public for other researchers at this https URL.

Deep Learning Models for Multilingual Hate Speech Detection

Machine learning techniques for hate speech classification of twitter data: State-of-the-art, future challenges and research directions

Analysis of subjective texts like offensive content or hate speech is a great challenge, especially regarding annotation process. Most of current annotation procedures are aimed at achieving a high level of agreement in order to generate a high quality reference source. However, the annotation guidelines for subjective content may restrict the annotators’ freedom of decision making. Motivated by a moderate annotation agreement in offensive content datasets, we hypothesize that personalized approaches to offensive content identification should be in place. Thus, we propose two novel perspectives of perception: group-based and individual. Using demographics of annotators as well as embeddings of their previous decisions (annotated texts), we are able to train multimodal models (including transformer-based) adjusted to personal or community profiles. Based on the agreement of individuals and groups, we experimentally showed that annotator group agreeability strongly correlates with offensive content recognition quality. The proposed personalized approaches enabled us to create models adaptable to personal user beliefs rather than to agreed offensiveness understanding. Overall, our individualized approaches to offensive content classification outperform classic data-centric methods that generalize offensiveness perception and it refers to all six tested models. Additionally, we developed requirements for annotation procedures, personalization and content processing to make the solutions human-centered.

Offensive, aggressive, and hate speech analysis: From data-centric to human-centered approach

Existing research on fairness evaluation of document classification models mainly uses synthetic monolingual data without ground truth for author demographic attributes. In this work, we assemble and publish a multilingual Twitter corpus for the task of hate speech detection with inferred four author demographic factors: age, country, gender and race/ethnicity. The corpus covers five languages: English, Italian, Polish, Portuguese and Spanish. We evaluate the inferred demographic labels with a crowdsourcing platform, Figure Eight. To examine factors that can cause biases, we take an empirical analysis of demographic predictability on the English corpus. We measure the performance of four popular document classifiers and evaluate the fairness and bias of the baseline classifiers on the author-level demographic attributes.

Multilingual Twitter Corpus and Baselines for Evaluating Demographic Bias in Hate Speech Recognition

Suicide is one of leading causes of death worldwide, yet little data is available about the lives of suicide victims because most people do not seek treatment. Research has shown that people express suicidal ideation in social media, which can potentially be tapped to improve our understanding of the thoughts and behaviors of people prior to suicide. In this work, we introduce a novel dataset of Chinese social media accounts of 130 people who committed suicide between 2011 and 2016. We describe the demographic and geographic composition of the users, then conduct a longitudinal text analysis of their post histories, showing observable changes in content leading up to the time of death. With encouraging exploratory findings, we discuss directions for future research.

Exploring Timelines of Confirmed Suicide Incidents Through Social Media

Low-dimensional vector representations of social media users can benefit applications like recommendation systems and user attribute inference Recent work has shown that user embeddings can be improved by combining different types of information, such as text and network data We propose a data augmentation method that allows novel feature types to be used within off-the-shelf embedding models Experimenting with the task of friend recommendation on a dataset of 5,019 Twitter users, we show that our approach can lead to substantial performance gains with the simple addition of network and geographic features

/pdf/incorporating-metadata-into-content-based-user-embeddings-5d48ohqxrn.pdf

Incorporating Metadata into Content-Based User Embeddings.

Bayesian inference methods for probabilistic topic models can quantify uncertainty in the parameters, which has primarily been used to increase the robustness of parameter estimates. In this work, we explore other rich information that can be obtained by analyzing the posterior distributions in topic models. Experimenting with latent Dirichlet allocation on two datasets, we propose ideas incorporating information about the posterior distributions at the topic level and at the word level. At the topic level, we propose a metric called topic stability that measures the variability of the topic parameters under the posterior. We show that this metric is correlated with human judgments of topic quality as well as with the consistency of topics appearing across multiple models. At the word level, we experiment with different methods for adjusting individual word probabilities within topics based on their uncertainty. Humans prefer words ranked by our adjusted estimates nearly twice as often when compared to the traditional approach. Finally, we describe how the ideas presented in this work could potentially applied to other predictive or exploratory models in future work.

Linzi Xing

Papers

Multilingual Twitter Corpus and Baselines for Evaluating Demographic Bias in Hate Speech Recognition

Exploring Timelines of Confirmed Suicide Incidents Through Social Media

Multilingual Twitter Corpus and Baselines for Evaluating Demographic Bias in Hate Speech Recognition

Incorporating Metadata into Content-Based User Embeddings.

Diagnosing and Improving Topic Models by Analyzing Posterior Variability.